Improved Genetic Algorithm Based Optimal Control for A Flying Inverted Pendulum

Abstract

We extended the inverted pendulum problem to the top of the quadrotor and evaluated the improved genetic algorithm (IGA). In order to solve the problem that it is difficult to determine the optimal weighting matrix in linear quadratic regulator (LQR). We optimized the IGA and the traditional genetic algorithm (GA) with the parameters of the LQR controller for different scenarios. We optimized the coding method, crossover operator and mutation rate in the IGA, and no longer randomly selected. The principle of optimal individual retention is adopted to ensure the convergence speed of the fitness value and the optimal solution that can be found. Compared with the effects of the optimized controller and the original controller, the results show that the IGA can effectively optimize the linear quadratic optimal control with better stability.